Ontogenetic Shift in Mouth Opening Mechanismsin a Catfish (Clariidae, Siluriformes):A Response to Increasing Functional DemandsD. Adriaens,1* P. Aerts,2 and W. Verraes1
1Ghent University, Vertebrate Morphology, Ghent, Belgium2University of Antwerp (UIA), Department of Biology, Antwerpen, Belgium
ABSTRACT During ontogeny, larval fish have to dealwith increasing nutritional and respiratory demands asthey grow. As early ontogeny is characterized by an in-creasing complexity of moving structural elements com-posing a fish skull, some constraints will have to be metwhen developing mechanisms, which enable feeding andrespiration, arise at a certain developmental stage. Thisarticle focuses on the presence/absence of a possible func-tional response in mouth opening during ontogeny inClarias gariepinus. Some reflections are given, based onmorphological data, as well as related function-analysisdata from the literature. Starting shortly after hatching, atotal of up to five different mouth opening mechanismsmay become functional. Of these, three may remain func-tional in the adult. As could be expected, the apparatusesthat enable these mechanisms show an increase in com-plexity, as well as a putative improvement in mouth open-ing capacity. Initially, two consecutive mechanisms mayallow a restricted depression of the lower jaw (both pas-
During their early life history, teleosts are contin-uously confronted with changing environmental fac-tors to which they have to respond in an adequatemanner in order to ensure their survival. As a con-sequence, the early ontogeny can be regarded as atrade-off between the race to respond to functionaldemands and the developmental process of Bauplandifferentiation (Galis et al., 1994; Osse et al., 1997).Events during the early life history of fishes, whichintroduce substantial changes in functional de-mands, involve respiration, feeding, or predatoravoidance (Otten, 1982, 1989, 1990; Osse and vanden Boogaart, 1995; Hunt von Herbing et al., 1996a;Fuiman, 1997; Osse et al., 1997).
The adult African catfish Clarias gariepinus(Burchell, 1822) (Siluriformes: Clariidae) has beensuggested to exhibit an opportunistic feeding behav-ior (Thomas, 1966). This is supported by morpholog-ical evidence of feeding structures, as well as stom-ach analyses (Groenewald, 1964; Teugels, 1986;Adriaens and Verraes, 1996, 1998; Adriaens et al.,1997). However, at the time C. gariepinus specimensreach the size when they can fully exploit their op-portunistic feeding behavior, they have survived a
larval phase featuring presumably less opportunis-tic feeding. Not only will the head size determine thekind of food that can be eaten, the size of the yolk sacand the structural differentiation of the feeding ap-paratus will be as restricting (Greenwood, 1955;Hecht and Appelbaum, 1987; Kohno et al., 1996b).Ontogenetic shifts in dietary demands are a commonfeature in teleosts (Thomas, 1966; Segnini and Bas-tardo, 1995; Olson, 1996; Lowe et al., 1996; Roweand Chisnall, 1996; Cook, 1996). In terms of “sym-morphosis” (i.e., the quantitative match between thefunctional capacity of an organism and the func-tional demands they have to cope with), it can beexpected that changes in diet and food particle sizeduring ontogeny will be coupled to changes in feed-ing mechanisms (Weibel and Taylor, 1981; Galis etal., 1994). Not only will the larva have to adapt to
Contract grant sponsors: the IWT and FWO; Contract grant num-ber: G.0388.00.
take up larger food items, it may also have to in-crease food capturing and manipulation efficiency(Osse, 1990). Larvae can do so by 1) making a com-plete use of those structures present, and/or 2) en-larging the capacity of these structures in responseto changes in demand (Galis et al., 1994). It is,however, crucial for a developing larva to do so at alow energy cost (Galis and de Jong, 1988) or tominimize energy loss during prey capture (Osse andDrost, 1989; Osse, 1990). Equally important is thetiming of the morphological transformations in rela-tion to the chronology of changes in functional de-mands and nutritional requirements (Otten, 1982;Holden and Bruton, 1994).
Apart from feeding, the efficiency of the respira-tory apparatus will have to increase as well, as gasexchange will no longer be sustained by cutaneousrespiration at a certain moment in a growing larva(Osse, 1989, 1990; Holden and Bruton, 1994).
In this study, it is hypothesized that such a rela-tion between functional demands and structural dif-ferentiations may exist in the cranial Bauplan aswell. In order to demonstrate the existence of such arelation, the ontogeny of mouth opening is investi-gated, with links being made with arising functionaldemands during early life history where possible.The functioning of possible mouth opening mecha-nisms is deduced from morphological data, but sup-port for the assumptions is based on kinematic datafrom the literature, as well as calculations of effi-ciencies. Although logical in an evolutionary con-text, however, most conclusions remain speculative.Some assumptions have been made previously (Ad-riaens and Verraes, 1994), however, on a generalbasis. This article presents a more detailed survey.
MATERIALS AND METHODSStudied Specimens
Several specimens of the African catfish Clariasgariepinus (Burchell, 1822), of different standardlengths were used (ranging between 4.1 mm SL (5standard length) (5 1 day posthatching) and 174.5mm SL (age unknown)) (see Adriaens and Verraes,1998: table 1). Eggs were obtained from the Labora-tory of Ecology and Aquaculture (Catholic Univer-sity of Leuven) and raised at a temperature of 25°C.The older juvenile specimens (100 days posthatch-ing) were commercially raised and obtained from W.Fleure (Someren, The Netherlands). Specimenswere sedated (using MS-222 overdose) and fixed in4% buffered formaldehyde of paraformaldehyde-glutaraldehyde solutions at different time intervals.Clearing and staining followed Hanken and Wass-ersug (1981), with trypsin being replaced by a 1%KOH solution. Cleared specimens were studied us-ing a stereoscopic microscope (WILD M5).
Several specimens were also used for serial sec-tioning in order to reveal details on myology andligaments. Embedding was done using Epon or
Paraplast. Sections of 2 and 5 mm, respectively, werestained with toluidine and an improved trichromestaining according to Mangakis et al. (1964). Sec-tions were studied using a Leitz Diaplan light mi-croscope. Drawings of both cleared and sectionedmaterial were made using a camera lucida. Three-dimensional reconstructions of serial sections weredone using a commercial software package (PC3D,Jandel Scientific, Sausalito, CA).
Mouth Opening Efficiencies
Serial sections of two specimens (5.6 and 7.2 mmSL) were used for calculating maximal muscle forcesas well as moments. Coordinates of muscle insertionsites were used for calculating the mean fiber length(as separate fibers could hardly be distinguished inthese early stages). Muscle volumes were calculatedusing the PC3D software in order to calculate thephysiological cross section area. Maximal outputforces (Fmax) of each muscle were calculated using26N/cm2 as the maximal force per unit of surface(Aerts, 1987; Aerts et al., 1987). Based on the lengthof the input arms (l), as well as the angle betweenthe input lever and the muscle (a), the moment ofeach muscle could be calculated (M 5 Fmax*sin(a)*l).
Data from cleared as well as sectioned specimensof several ontogenetic stages of Clarias gariepinus(7.2, 19.0, 21.5, 41.9, 127.0, and 147.7 mm SL) wereused to calculate efficiencies of the opercular andhyoid four-bar systems using the model of Aerts andVerraes (1984). This model was also modified inorder to calculate the efficiencies of both hyoid andopercular systems at specific angles of mandibulardepression, especially output velocities. This dimen-sionless parameter gives an idea of the difference ininput speed of the crank, compared to the outputspeed of the follower of a four-bar system, takinginto account the lengths of the input and outputlevers. The aims of this modeling are: 1) to distin-guish which ontogenetic pattern in efficiency alter-ations is present; 2) to compare the efficiency of theopercular four-bar system with that of the hyoidfour-bar system in several ontogenetic stages; 3) tocheck to what degree the hyoid and opercular bar-systems optimize mouth opening; and 4) to comparethe efficiency of the opercular four-bar system inseveral ontogenetic stages with that of other teleosts(data obtained from Aerts and Verraes, 1984) (seeDiscussion). Data from the same specimens werealso used for calculating the degree of suspensorialabduction during hyoid depression, based on themodel of the de Visser and Barel (1996).
RESULTSMouth Opening Mechanisms
The detailed morphology of cranial ontogeny hasbeen dealt with in previous articles: chondrocra-nium (Adriaens and Verraes, 1997a,e), osteocra-
198 D. ADRIAENS ET AL.
nium (Adriaens and Verraes, 1998), and myology(Surlemont et al., 1989; Surlemont and Vandewalle,1991; Adriaens and Verraes, 1996, 1997b,c,d). Con-sequently, only those features needed to understandthe working of the mouth opening apparatus will begiven here. Based on the morphology of the speci-mens used, a total of five possible mouth openingmechanisms could be distinguished. The results arepresented according to the expected chronologicalorder. For each mechanism, the morphology anddeduced possible functioning are described.
Mouth opening probably only occurs at about 5.0mm SL (corresponding to 48 h posthatching). Thepresence of the adductor mandibulae complex, theposterior intermandibular, and anterior interhyoid
muscles has been observed in a 4.7 mm TL speci-men; however, no movements of the lower jaw werepresent as their insertions were still lacking (Sur-lemont and Vandewalle, 1991). Surprisingly, how-ever, the figures in this article suggest the presenceof insertions (as no sectioned material of comparablestages was at hand, this could not be verified).
Mechanism 1 (5.0–5.6 mm SL)Form. Both form and function of this mechanism
have been described by Surlemont et al. (1989). Athatching, the suspensorium, lower jaw, and hyoidbar arise as a single cartilaginous plate (Fig. 1A).Both lower jaw and hyoid bar are connected to the
Fig. 1. Morphology of mouth opening apparatuses in Clarias gariepinus during ontogeny. A: Apparatus 1: neurocranium, suspen-sorium, lower jaw, and hyoid bar of the 5.0 mm SL stage (modified from Surlemont et al., 1989). B: Apparatus 2: graphical3D-reconstruction of the same structures, without the neurocranium in the 5.6 mm SL stage. C: Apparatus 3: graphical 3D-reconstruction of the same structures in the 7.2 mm SL stage. D: Apparatus 4: same as C. ch, ceratohyale; c-Meck, cartilago Meckeli;hs, hyosymplecticum; ih, interhyale; l-mh, ligamentum mandibulo-hyoideum; m-A2A3’, musculus adductor mandibulae A2A3’; m-A3“,musculus adductor mandibulae A3”; m-inth-a, musculus interhyoideus anterior; m-intm-p, musculus intermandibularis posterior;m-l-ap, musculus levator arcus palatini; m-pr-h, musculus protractor hyoidei; m-sh, musculus sternohyoideus; o-cl, os cleithrum;ot-cap, otic capsule; pal, palatinum; p-q, pars quadrata of the pterygoquadratum; prc-co, processus coronoideus; prc-op, processusopercularis of the hyosymplecticum; prc-pt, processus pterygoideus of the pterygoquadratum; prc-ra, processus retroarticularis;tn-m-p, taenia marginalis posterior; tr-cr, trabecula cranii.
199MOUTH OPENING MECHANISMS IN A CATFISH
suspensorium by a narrow, pliable region. Shortlyafter hatching (at about 5.2 mm TL), the adductormandibulae, the posterior intermandibular, and an-terior interhyoid muscles become attached to boththeir insertion sites (Surlemont et al., 1989). In theirarticle, Surlemont et al. (1989) surprisingly con-cluded that “as the muscles likely to lower the man-dible are not yet inserted on the skeleton, the lowerjaw could come back to its resting position by its ownelasticity,” after having mentioned the contradictorystatement that “some muscles are provided withboth proximal and distal insertions:…, the protrac-tor hyoidei,…” The insertion of this protractor,which at that stage still consists of the distinct in-termandibularis posterior and interhyoideus ante-rior, on both the lower jaw and the ceratohyal couldbe observed in a 5.8 mm TL specimen, in which thelower jaw still appeared to be continuous with thesuspensorium.
Function. If indeed the posterior intermandibu-lar and anterior interhyoid muscle are attached tothe lower jaw and the ceratohyal in this early stage,mouth opening may be more efficient, as proposed bySurlemont et al. (1989). It would imply that mouthopening could occur passively as well as throughmuscular contraction. The passive mechanismwould rely on the elastic properties of the cartilagi-nous connection between lower jaw and suspenso-rium, as strain energy stored in that connectionduring adductor mandibulae contraction would re-store the mouth to its original, slightly opened posi-tion during adductor relaxation (Fig. 2A). The activemechanisms would involve the muscular retractionof the lower jaw by the posterior intermandibularisand anterior interhyoideus muscles. However, somedrawbacks have to be considered: 1) in order for thepassive mechanism to work, the strain energy re-leased during adductor mandibulae relaxation mustgenerate a mouth opening force which exceeds theresistive force of the extending adductor muscle fi-bers, as well as that needed to displace the watermass below the lower jaw. As a consequence, maxi-mal mouth opening in this mechanism could bereached very quickly, thus enabling very littlemouth opening. 2) Further mouth opening couldthen be enabled due to the active mechanism involv-ing the muscles interconnecting the lower jaw withthe hyoid bar. However, at this 5.2 mm TL stage, thesternohyoideus muscle (needed for fixation of thehyoid bar) is still unattached to the hyoid bar (Sur-lemont et al., 1989), which could thus annul anymouth opening action. Even then, arguments can bemade that suggest a possible (but restricted) mouthopening action of this mechanism, as described inMechanism 2.
Mechanism 2 (5.6–19.2 mm SL)Form. At approximately 5.6 mm SL, Meckel’s car-
tilage becomes detached from the suspensorium and
an articulation is formed. The hyoid bar is still con-tinuous with the suspensorium through a cartilagi-nous interhyal. The posterior intermandibular andanterior interhyoid muscles still form a distinctX-shaped muscle pair, interconnecting the lower jawwith the hyoid bar (Fig. 1B).
Function. Once a true articulation is formed be-tween the lower jaw and the suspensorium, strainenergy can no longer be stored, thus annulling thepreviously mentioned passive mechanism. However,a true articulation also implies a substantial reduc-tion of resistive forces in the connection betweenlower jaw and suspensorium during mouth opening.This would improve the mouth opening throughmuscular contraction (Fig. 2B). However, theoreti-cally, contraction of the posterior intermandibularand anterior interhyoid muscle complex may notonly result in a lower jaw depression, but in a hyoidelevation as well. As the sternohyoideus cannot pre-vent this action (due to the absence of its insertion),this action could thus partially nullify mouth open-ing. Although the differences in stresses in the man-dibular and hyoid connections with the suspenso-rium alone could allow mouth opening between 5.6and 6.2 mm SL, unfavorable muscle–skeleton con-figurations might overrule this. Differences in mo-ment about the mandibular articulation and thehyoid connection with the suspensorium can thusmake this mechanism more effective or may over-rule the resistive forces. In the 5.6 mm SL specimen,the calculated moments, which could be generatedby the intermandibular and interhyoid musclesabout the interhyal, are much lower than that aboutthe mandibular articulation (21.17 and 32.59mN.mm, respectively). In the 7.2 mm SL specimen,this calculated difference is substantially greater(66.18 and 222.59 mN.mm, respectively). One has tobear in mind that these moments represent a staticsituation, which is not the case in a dynamicallymouth-opening larva. The only moments that couldprovide a conclusive idea of the efficiency of themouth opening mechanism is that at the startingposition, i.e., with the mouth closed. In the exam-ined specimens, however, the mouths were partiallyopen (gape of about 30° and 20° in the 5.6 and 7.2mm SL specimens, respectively), so some precautionhas to be taken in considering these results.
Theoretically, this mechanism could remain func-tional until the interhyal becomes detached from thehyoid bar (or the suspensorium), which occurs atabout 19.2 mm SL (personal observations andVandewalle et al., 1985).
Mechanism 3 (6.2–… mm SL)Form. At 6.2 mm SL, the articulatory facet of the
lower jaw has become more pronounced, eventhough ossifications surrounding this facet are stilllacking (Fig. 1C). Posterior to the articulation,Meckel’s cartilage has formed a pronounced retroar-
200 D. ADRIAENS ET AL.
ticular process to which the mandibulo-hyoid liga-ment is attached. Later during ontogeny, this liga-ment becomes a stout connective tissue string, lying
medial to the equally sized angulo-interopercularligament. At 6.2 mm SL, the sternohyoideus muscleis connected to the hypohyal cartilage, as well as to
the cleithral bone, and may thus become functional.In the juvenile stage, the sternohyoideus is a verylarge muscle, being especially broad posteriorly (Ad-riaens and Verraes, 1997b: fig. 8C). Rostrally thismuscle inserts onto the forked parurohyal bone,which attaches through a short, paired ligament tothe ventral hypohyal ossification (Adriaens and Ver-raes, 1998: fig. 4).
Function. From 6.2 mm SL on, mouth openingcan now presumably occur more powerfully. Thepresence of the mandibulo-hyoid ligament (ligamen-tum angulo-ceratohyale of Adriaens and Verraes,1997b) enables the coupling of hyoid depression dur-ing sternohyoideus activity to the jaw depression(Fig. 2C). This hyoid mouth opening mechanismworks as a four-bar system (Fig. 3A). In this four-barsystem the following constituents are present: 1) theframe, the bar between the mandibular articulationand the interhyal-suspensorium connection; 2) thecrank, the bar between interhyal-suspensorium con-nection and insertion site of the mandibulo-hyoidligament on the hyoid; 3) the coupler, themandibulo-hyoid ligament; and 4) the follower, thebar between the mandibular articulation and theinsertion site of the latter ligament on the retroar-ticular process. The depression of the hyoid bars canbe observed in a 6.2 mm SL larva (6.8 mm larva ofSurlemont and Vandewalle, 1991). The maximalgape (about 40° in the 7.2 mm SL stage) would bereached once the working line through the ligamentreaches the mandibular articulation (Fig. 2C). Incase neurocranial elevation would occur, this couldspeed up mouth opening, as well as make it morepowerful. However, maximal gape would not be af-fected.
Mechanism 4 (6.2 mm SL–…)Form. In the 6.2 mm SL stage, the intermandibu-
laris posterior and interhyoideus anterior musclescan hardly be distinguished, as they have fused toform the protractor hyoidei (Fig. 1D). The lattermuscle interconnects the rostral part of the lowerjaw with the posterolateral part of the hyoid bars, aswell as it embeds the mandibular barbel bases (Ad-riaens and Verraes, 1997b). The sternohyoideusforms the connection between the cleithral bone andthe hyoid bars, whereas hypaxial muscles insertonto the posterior margin of the latter bone.
Function. According to this mechanism, the de-pression of the lower jaw would be coupled muscu-larly to the depression of the hyoid bars. In this case,the pectoral retraction (or fixation) through the con-traction of the hypaxials, and the depression (orretraction) of the hyoid bars through the contractionof the sternohyoideus, could assist the mandibulardepression through the contraction of the protractorhyoidei (Fig. 2D) (Osse, 1969), provided that its lineof action is situated below the jaw suspension. Thelatter is the case in all stages of Clarias gariepinus
(Fig. 5). Consequently, this mechanism could allowfurther mouth opening, once maximal gape is ob-tained in Mechanism 3.
Mechanism 5 (11.1 mm SL–…)Form. At about 11 mm SL, the majority of cranial
bones have formed in Clarias gariepinus (Adriaensand Verraes, 1998: fig. 25). The crucial, newlyformed structures for this mouth opening mecha-
Fig. 3. Scheme of the hyoid and opercular four-bar systems inClarias gariepinus. A: Hyoid system. B: Opercular system.C: Both systems transposed (Fin, input force, Fout, output force).
202 D. ADRIAENS ET AL.
nism involve the interopercular bone and the corre-sponding ligamentous connections with the lowerjaw and the opercular bone. In the 11.1 mm SLlarva, a rudimentary interopercular bone could beobserved, with a ligamentous connection to both theretroarticular process and the opercular bone (incontrast to the observations of Vandewalle et al.,1985, where this ligament was only observed at 21.5mm SL) (Fig. 4A). The lower jaw now bears all itsossifications (with exception of the coronomeckelianbone). Later during ontogeny, the ligament becomeslarger, whereas the retroarticular process bears asmall process for its attachment (Adriaens and Ver-raes, 1998: fig. 20A). The levator operculi musclemay become functional in Clarias gariepinus atabout 6.8 mm SL (Adriaens and Verraes, 1997b:table 1), whereas in juveniles it inserts on the com-plete dorsal margin of the opercular bone (Fig. 4B).The opercular bone is now tilted substantially (in aclockwise direction, for a lateral view of the leftside), compared to the larval stages.
Function. The opercular four-bar system is awell-known mouth opening mechanism in teleosts(Aerts et al., 1985; Lauder and Liem, 1989). Theinput force is generated by the contraction of theopercular levator, making it rotate (Fig. 2E). Theforce is transmitted through the interopercular boneand both the ligaments attached to it, consequentlypulling onto the retroarticular process of the lowerjaw (Fig. 3B). The lower jaw is depressed, the max-imum gape being reached once the axis through theinteropercular bone and its ligaments runs throughthe mandibular articulation (about 35° in the 19.1mm SL stage). Based on the changes in the relativelength of the four bars, ontogenetic changes in theefficiency of this mouth opening mechanism can beexpected (Fig. 4) (see Discussion).
Mouth Opening Efficiencies
Mouth opening can only occur when the momentsof the mouth opening muscles exceed that of coun-teracting forces, such as 1) the momentum of thewater which has to be displaced, 2) the inertialforces of the accelerating jaw, 3) the negative pres-sure in the oral cavity, which is generated duringmouth opening, and 4) the strain energy stored inmouth closing muscles that have relaxed (Aerts etal., 1987). It is thus of the utmost importance thatmouth opening mechanisms are sufficiently efficientto overcome these forces.
In order to get any idea of the efficiency of mouthopening, calculating output velocities of the four-barsystems can provide some information. During on-togeny, the lengths of the bars, in relation to theframe, seem to change allometrically in Clariasgariepinus, which suggests a difference in efficiency.The ratio of the crank and follower appears to behigher in the opercular system, which would suggest
that the opercular system is kinematically more ef-ficient than the hyoid one. When taking into accountthe ratio of output and input lever lengths, the dif-ference in output velocities will even be higher. Themodel supports these findings: 1) the output velocityof the hyoid system is larger in large juvenile stages,compared to larval stages; 2) the opercular system iskinematically more efficient than the hyoid system(Fig. 6). The efficiency of the opercular system in C.gariepinus, compared with that of some teleosts,shows that the output velocity of the opercular four-bar system in Clarias is substantially lower thanthat of the some other teleosts, and thus much moreforce efficient (Aerts and Verraes, 1984). The factthat the output force efficiency of the opercular four-bar system in Clarias is high does not imply thatlarge forces will be generated on the lower jaw, asthe size of the input force (from the levator operculi)is crucial as well. The question as to what degree theopercular four-bar system participates in generatingthe required mouth opening force remains unan-swered (EMG studies are in progress).
The question can be raised as to what degree thehyoid and opercular four-bar systems contribute tothe efficiency of mouth opening, and how thischanges during ontogeny. The output velocity of thehyoid and opercular four-bar systems during mouthopening appears to show some kind of pattern, al-though the 19.0 mm SL stage does not always fit it.1) The range of mandibular angles, during which theoutput velocity of the opercular system has not yetreached an arbitrary level of 50, increases duringontogeny (Fig. 7). The lowest range is observed inthe 21.5 mm SL stage (621°), whereas in the 147.7mm SL stage it reaches approximately 37°. The val-ues for the hyoid system remain fairly constant be-tween 25° and 30°. 2) The difference between outputvelocity values of the hyoid and opercular system ishigher in early stages compared to later stages. Thelow levels of the hyoid output velocity suggest ahigher output force, which is especially importantsince the input force in this four-bar system is muchhigher as well, compared to the opercular four-barsystem. 3) During early ontogeny, the opercular out-put velocity reaches an arbitrary level of 50 equallyas fast or even faster than the hyoid one. In laterstages, the hyoid output velocity reaches that levelmore rapidly than does the opercular output veloc-ity. Theoretically, it can be said that if input forceswere equal the hyoid four-bar system would be moreimportant for generating a powerful mouth opening,especially during early ontogeny, whereas the oper-cular system would enable a fast mouth opening.However, this would be a simplification to a largedegree, as 1) input forces differ largely (small levatoroperculi and large sternohyoideus), and 2) bothmechanisms will act and interact in a more complexmanner (Fig. 3C).
203MOUTH OPENING MECHANISMS IN A CATFISH
Fig. 4. Opercular four-bar system (Mechanism 5) in Clarias gariepinus.. A: 11.6 mm SL stage.B: 125.5 mm SL. m-l-op, musculus levator operculi; mnd, mandibula; o-hm, os hyomandibulare;o-iop, os interoperculare; o-op, os operculare; o-q, os quadratum.
204 D. ADRIAENS ET AL.
DISCUSSIONOntogenetic Shift in Mouth OpeningMechanisms
Mouth opening in teleosts is establishedthrough a wide range of mechanisms, from verysimple ones to a whole complex of integrated cou-plings. Simple mechanisms involve the coupling ofhyoid depression to mandibular depression (Ot-ten, 1982; De La Hoz and Aldunate, 1994; Huntvon Herbing et al., 1996a). Mouth opening be-comes more complex once four-bar systems areinvolved (Anker, 1974; Aerts and Verraes, 1984;Westneat, 1990, 1994). More complex mouth open-ing mechanisms involve the coupling of the man-dibular depression to a whole set of other mecha-nisms: neurocranial elevation (Muller, 1987),intracranial angulation (Lauder, 1980b), maxil-lary rotation (Aerts and Verraes, 1987; Westneatand Wainwright, 1989; Westneat, 1990), premax-illary protrusion (Westneat and Wainwright,
1989; Westneat, 1990), and even lower jaw protru-sion (Westneat and Wainwright, 1989).
As the complexity of a mechanism can be relatedto the number of elements involved, it can be ex-pected that the complexity of mouth opening mech-anisms may increase during ontogeny. This impliesthat an ontogenetic shift in mouth opening mecha-nisms is not unfavorable in order to constantly op-timize mouth opening (in relation to the apparatuspresent). Such a shift appears to be present in Clar-ias gariepinus, where an increase in complexity isapparent (Fig. 8). Ontogenetic shifts in mouth open-ing mechanisms have been noted in other fishes aswell (Verraes, 1977; Otten, 1982; Liem, 1991; Huntvon Herbing et al., 1996a; Hunt von Herbing, 1997).In general, two types of mouth opening mechanismsare distinguished in teleosts: 1) a hyoid mouth open-ing mechanism, and 2) an opercular mouth openingmechanism. Passive mouth opening has been intro-duced by Surlemont et al. (1989).
Fig. 5. Scheme of the functional shift of the protractor hyoidei.
205MOUTH OPENING MECHANISMS IN A CATFISH
Passive Mouth Opening Mechanisms
The presence of passive mouth opening mecha-nisms has generally been overlooked, as it presum-ably only plays a crucial role in early life stages, ifpresent. In the case of Clarias gariepinus, Surle-mont et al. (1989) emphasized the importance of thecartilaginous connection between the lower jaw andthe suspensorium in mouth opening. However, asmuscular action cannot be excluded at this stage,the importance of passive mouth opening should notbe overrated. It may play a supportive role in mouthopening, as at this stage the muscles are very rudi-mentary. A study on the contraction capacities ofsuch primordial muscle tissue could provide someinformation on the necessity for such a passivemouth opening mechanism during early ontogeny.
If, indeed, as suggested by Surlemont et al. (1989),muscle contraction of the ventral muscles cannot
induce mouth opening, this passive mechanism doesplay a crucial role, but only until a true articulatoryfacet is formed on the lower jaw. Although feedingstill occurs endogenously at this stage, respiratoryrequirements for buccal ventilation may increase.
Hyoid Mechanisms
The hyoid mechanism involves the ligamentouscoupling of hyoid depression to mouth opening. Al-though the presence of such a mandibulo-hyoid lig-ament in teleosts has been linked to lower jawlength (Verraes, 1977), its presence appears to be aplesiomorphic feature of non-ctenosquamate te-leosts, as well as lower Actinopterygii (Lauder,1980a; Johnson, 1992; Stiassny, 1996). In aulopi-form fishes, this ligament becomes extended poste-riorly, thus inserting onto the interopercular bone,
Fig. 6. Output velocity ofthe hyoid and opercular four-bar system during ontogenyin Clarias gariepinus.
206 D. ADRIAENS ET AL.
whereas in Ctenosquamata the connection with thelower jaw is lost. In several non-ctenosquamate te-leosts, the ligament can be observed early duringontogeny: Clarias gariepinus, Oncorhynchus mykissWalbaum, 1792, and Gadus morhua (Verraes, 1977;Hunt von Herbing et al., 1996a; Adriaens and Ver-raes, 1997b).
As mentioned in Results, the role of the protractorhyoidei as a mouth opener also depends on the po-sition of its line of action in relation to the mandib-ular joint (Otten, 1982) (Fig. 5). If ventral to thisjoint, the protractor will enable mouth opening.However, once dorsal to the articulation the functionof the muscle shifts from mouth opening to mouth
Fig. 7. Output velocities of hyoid and opercular four-bar systems during mouth opening inClarias gariepinus. A: 19.0 mm SL. B: 21.5 mm SL. C: 41.9 mm SL. D: 127.0 mm SL. E: 147.7mm SL.
207MOUTH OPENING MECHANISMS IN A CATFISH
closing, as observed in a cichlid species (Fig. 5)(Anker, 1974; Elshoud-Oldenhave and Osse, 1976).Such shifts are also observed in Pomacentridae andEmbiotocidae, whereas in the latter the shift occursduring the intraovarian life stage (Liem, 1991). Asin cod the protractor hyoidei remains a mouthcloser, the presence of a mandibulo-hyoid ligament
is crucial for mouth opening during early life stages(Hunt von Herbing et al., 1996a). In Clarias gariepi-nus the protractor remains a mouth opener duringthe entire ontogenetic period, which means that,theoretically, both the ligamentous and the muscu-lar hyoid mouth opening mechanisms could gener-ate mouth opening. Supportive evidence comes from
Fig. 8. Overview of the morphological features, with corresponding mouth opening mechanisms in Clarias gariepinus. l-an-iop,ligamentum angulo-interoperculare; l-mh, ligamentum mandibulo-hyoideum; l-op-iop, ligamentum operculo-interoperculare; m-ad-mnd, musculus adductor mandibulae complex; m-inth-a, musculus interhyoideus anterior; m-intm-p, musculus intermandibularisposterior; m-l-op, musculus levator operculi; m-pr-h, musculus protractor hyoidei; m-sh, musculus sternohyoideus; o-iop, os intero-perculare.
208 D. ADRIAENS ET AL.
the ontogenetic change in the quadrate angle (asdepicted in the inset of Fig. 10). This angle gives anindication of the ventral and anterior displacementof the mandibular joint during ontogeny (large an-gles assume a mandibular joint below the protractorworking line) (Otten, 1982; Hunt von Herbing et al.,1996a). In Astatotilapia elegans (Trewavas, 1933),this angle increases rapidly (up to 90°), where at the6 days critical period the angle increases to a muchlesser degree when the protractor becomes a mouthcloser (Fig. 5). A large quadrate angle (up to about160°) is observed during the entire ontogenetic pe-riod in Gadus, whereas in Clarias and Oncorhyn-chus the angles remain fairly low. The dorsoven-trally flattening of the skull can partially explain thelow values of quadrate angles in C. gariepinus. Sometrends can be observed: 1) low quadrate angles areobserved at the moment the yolk sac is resorbed(except for Astatotilapia); 2) maximal quadrate an-gles are obtained at the moment the opercular four-bar system becomes functional (except for Astatoti-lapia and Micrometrus). This may suggest that themuscular hyoid apparatus may play an importantrole in mouth opening during early ontogeny, espe-cially for respiratory purposes, whereas mouthopening for feeding requires the involvement ofother mechanisms.
Kinematically, both hyoid mechanisms (i.e., theligamentous and muscular) can hardly be distin-guished from each other. However, it has been sug-gested that the ligamentous hyoid mechanism in-volves a simultaneous depression of the hyoid bar
and the lower jaw (Hunt von Herbing et al., 1996a),whereas in the muscular hyoid mechanism the hyoiddepression follows the mandibular one shortly after(Otten, 1982). Data from EMG support this idea(Lauder, 1980a). In 6.2 mm SL larvae, the hyoiddepression was observed to initiate shortly after themandibular depression, which might indicate thatthe muscular hyoid mechanism is responsible formouth opening (Fig. 9) (Surlemont and Vandewalle,1991). This, however, brings up the question con-cerning the functioning of the mandibulo-hyoid lig-ament (see below).
The synchronous action of neurocranial elevationduring mouth opening has been observed in severalteleosts (e.g., Osse, 1969; Lauder, 1979; Muller,1987). However, in Clarias, as well as in catfishes ingeneral, the highly modified Weberian apparatusbecomes increasingly large and broad during ontog-eny, as well as the vertebrae of the complex becom-ing fused to each other (Radermaker et al., 1989).Consequently, the spatial constraints of the nuchalarea may restrict the effect of neurocranial eleva-tion. However, this would only imply that the rota-tion point becomes shifted backward, between thevertebral complex and the sixth vertebra. As ob-served in cichlids, neurocranial elevation is not re-stricted to a focused rotation point between two ver-tebrae, but rather a bending of the anterior part ofthe vertebral column (Aerts, 1987). A supraoccipitalcrest, which is generally associated with improvedneurocranial elevation, is absent in Clarias. How-ever, at the ventral side of the parieto-supraoccipital
Fig. 9. Kinematics of mandibular and hyoid depression in larval Clarias gariepinus (6.2 mmSL), showing the time lapse between the onset of both depression movements. LJ, lower jaw; HB,hyoid bar. (Modified from Surlemont and Vandewalle, 1991.)
209MOUTH OPENING MECHANISMS IN A CATFISH
bone complex, a large vertical crest is present, con-nected to the Weberian apparatus (Adriaens andVerraes, 1998: fig. 16). This crest provides a goodalternative for an enlarged insertion surface andlever action for the epaxial muscles, fitting into thedesign of the dorsoventrally flattened skull. Neuro-cranial elevation has been observed during feedingin Ictalurus (Ictaluridae) (Alexander, 1970).
Opercular Mechanisms
The opercular mechanism is observed in mostadult teleosts, where it functionally replaces or as-sists the hyoid mouth opening mechanism (Anker,1974; Lauder, 1982; Otten, 1982; Lauder and Liem,1989; Liem, 1991; De La Hoz and Aldunate, 1994).The contraction of the levator operculi during mouthopening has been demonstrated by numerous elec-tromyographic studies (Osse, 1969; Elshoud, 1978;Lauder and Liem, 1980; Lauder, 1981; Liem, 1984;Wainwright and Turingan, 1993). In some teleosts(e.g., Cichlidae, Pomacentridae), the opercularmechanism becomes functional synchronously withthe functional shift of the protractor hyoidei frommouth opening to mouth closing (Otten, 1982; Liem,1991) (Figs. 5, 10). In these teleosts, the synchroni-zation of this shift in mouth opening mechanisms isthus crucial for their survival. The shift from the
hyoid mechanism to the opercular mechanism oc-curs very early during ontogeny in those species thatlack a mandibulo-hyoid ligament (after 6% of thelarval period) (Hunt von Herbing et al., 1996a). Ingeneral, the opercular mechanism becomes func-tional after the yolk sac has been resorbed (Fig. 10),which may suggest that the shift may be related toan ontogenetic shift in feeding behavior (see below).Cichlids seem to differ from this general trend,which can be explained by the fact that very earlyduring ontogeny (i.e., during the yolk sac stage) boththe ligamentous and muscular hyoid mechanismscannot function (both the mandibulo-hyoid andhyoid-interopercular ligaments are missing and theprotractor hyoidei is a mouth closer, respectively).The functioning of the opercular mechanism is thuscrucial. In Embiotocidae such an acute shift ispresent as well: the opercular system becomes func-tional at the moment the protractor becomes amouth closer (the ligamentous hyoid system cannotfunction due to the absence of the ligament), al-though well after the yolk sac has been resorbed(Fig. 10).
Although in some teleosts an acute shift occursfrom one mouth opening mechanism to another, sev-eral examples can be given in which a clear overlapoccurs (Lauder and Liem, 1989), including Clariasgariepinus (Figs. 8, 11). Between 6.2 and 11.1 mm
Fig. 10. Graph of the quadrate angle during ontogeny in five teleosts (data based on Otten,1982; Hunt von Herbing et al., 1996a, Liem, 1991).
210 D. ADRIAENS ET AL.
SL, both hyoid mouth opening mechanisms (Mech-anisms 3 and 4) may be functional (the effect of thecartilaginous interhyal in Mechanism 2 is neglectedhere). From 11.1 mm SL on, the opercular mecha-nism may become functional as well, thus leavingthree mechanisms that can enable mouth opening atthe same moment. An overlap of mechanisms be-comes advantageous if protagonism exists, with achronological shift of their specific efficiency peaksduring the process of mouth opening, as suggestedfrom the model: during the earlier stages the oper-cular mechanism appears to be velocity-efficient,whereas the hyoid mechanism appears to be moreforce-efficient (Fig. 7). Opercular rotation coupled tomandibular depression has frequently been ob-served; however, maximal gape is not alwaysreached when the opercular rotation is maximal
(Westneat, 1990). This suggests that the opercularmouth opening mechanism is only functional duringa certain phase of mouth opening, and in order toobtain a maximal gape another mechanism has totake over. It has been suggested in cichlids that therole of the opercular mouth opening mechanism dur-ing feeding involves the triggering of mouth open-ing, whereas the mandibular depression is exertedby the contraction of the hyoid musculature (Lauder,1981; Aerts et al., 1987). Ontogenetically protago-nist mechanisms may thus chronologically be pro-tagonists as well, meaning that they may assist eachother during mouth opening in keeping up the re-quired efficiency at all levels of mouth gape (seebelow).
Overlap of mechanisms may also allow a shift infunction in one of the mechanisms (Lauder et al.,
Fig. 11. Overview of the ontogenetic shift in mouth opening mechanisms in relation to func-tional demands: Phase 1 5 passive, aquatic, cutaneous respiration; Phase 2 5 active, aquatic, gillrespiration; Phase 3 5 active (exogenous) feeding; Phase 4 5 carnivory; Phase 5 5 aerialrespiration.
211MOUTH OPENING MECHANISMS IN A CATFISH
1989; Lauder and Liem, 1989). It may thus well bethat the simultaneous action of the muscular andligamentous hyoid mechanisms has allowed the dif-ferentiation of the protractor hyoidei in a superficialpart controlling the mandibular barbels, and adeeper part, which presumably still assists in mouthopening (Ghiot et al., 1984; Adriaens and Verraes,1997b). This would suggest that the ligamentousmouth opening mechanism would be of greater im-portance compared to the muscular system, thuscontradicting what has previously been said. How-ever, to clarify this an electromyographical analysisof larval Clarias gariepinus mouth opening would berequired in order to determine to what degree theprotractor hyoidei assists in mouth opening. Anoverlap of mouth opening mechanisms may haveenabled the structural innovations observed in Lori-cariidae and Astroblepidae, where the connectionbetween the lower jaw and the opercular bone is lost(Schaefer, 1987; Schaefer and Lauder, 1996).
Ontogenetic Shift in Feeding Behavior
During ontogeny, fishes increase their prey cap-ture success. Consequently, mouth opening mayhave to be improved in order to occur faster (Cough-lin, 1994; Cook, 1996). As it has been estimated thata prey size of 0.6 times a fish mouth size gives thegreatest energetic benefit in relation to time cost,the gape size will be important also (Gill, 1997).Changes in feeding mode generally involve changesin ram- or suction-actions during feeding, which arethe result of ontogenetic alterations in morphologyand behavior (Cook, 1996). In Lates calcarifer (Cen-tropomidae), the capture efficiency in early larvae(i.e., 10–20 h posthatching) largely depends on thesize of the larvae, whereas in older larvae (i.e.,60–70 h posthatching) the feeding ability is greatlyimproved by the ontogenetic differentiations of theskull. At that stage, structures have formed thatallow suction/grasping and manipulation of prey,whereas earlier prey capturing is dependent on suc-tion abilities only (Kohno et al., 1996b). Because ofthe fact that fishes live in “an environment with adensity 800 times and viscosity 30 times that of air,”they require a highly effective mechanism for preycapture (Osse, 1990). This could be done by produc-ing a high forward thrust and some suction feedingin order to engulf prey, or by producing a high suc-tion force. However, as Osse (1990) stated, thatwhen “larvae start external feeding, their suctionforces must be high because little contribution ofswimming and none of protrusion is available,”many fish larvae start as suction feeders (Osse,1990; Kohno et al., 1996a). In carp larvae of 6.5 mm,a negative pressure inside the orobranchial cavityhas been estimated to reach –300 Pa (Drost et al.,1988). Suction feeding has been observed in otherteleostean larvae (Coughlin, 1994), although more
straining-like feeding types have been suggested aswell (Kohno et al., 1996a).
One crucial factor for suction feeding is the pres-ence of an opercular cover (Osse, 1990). This impliesthat, for cod, suction feeding would only be possibleafter 3-4 weeks posthatching (about 6 mm larvae),as the opercular bone fails to develop until then(Hunt von Herbing et al., 1996b). In Clarias gariepi-nus, the opercular bone is the first bone to develop(formation starts at about 4.1 mm SL, 1 day post-hatching). Initial orobranchial expansions are theo-retically possible from 5.0 mm SL on (Surlemont etal., 1989). These involve the abduction of the sus-pensoria, the elevation of the neurocranium, and thedepression of the lower jaw (Adriaens and Verraes,1997d: table 1). Functional teeth are observedshortly after (at about 6.0 mm SL). At that stage,however, larvae still possess a yolk sac, implyingthat active feeding may not be crucial yet. It may bepossible that suction action at that time is mainlyfor respiratory purposes.
The application of the model of de Visser andBarel (1996) to the data of Clarias gariepinus sug-gested that in early larval stages the sternohyoideuscontraction will generate substantial hyoid depres-sion but little suspensorial abduction, with a grad-ual shift to increasing suspensorial abduction dur-ing later ontogeny. The reduction of hyoiddepression in juvenile stages could be related to thedorsoventral flattening of the skull. In a way, itcould explain the ontogenetic shift from a cartilagi-nous connection to a ligamentous connection be-tween the suspensorium and the interhyal: the ac-tion onto the hyoid bar can consequently berestricted to a retraction instead of a substantialdepression (Adriaens and Verraes, 1994). The reduc-tion of substantial hyoid depression in juvenile C.gariepinus fits into the previously formulated hy-pothesis that such a depression is not required andwould even be disadvantageous for dorsoventrallyflattened species, like C. gariepinus (Adriaens andVerraes, 1994, 1997e). Its major role may then be formouth opening, as it pulls on the lower jaw whenretracted. However, true evidence to support or in-validate this hypothesis has to come from kinematicanalysis (in progress). A sternohyoideus below thehorizontal hyoid plane and low da-values, as well asfeeding behavior (Hecht and Appelbaum, 1987), sug-gest that during early stages C. gariepinus larvaeperform a suction/grasping way of feeding, whereaslater juvenile stages are piscivore biters. This canalso be derived from the overall skull morphology.Suction feeding is improved if the skull changesfrom tube-like to cone-like, implying pointed snoutsand high heads (Alexander, 1965, 1967; Muller andOsse, 1984; Hunt von Herbing et al., 1996a). In C.gariepinus, skull morphology approaches this situa-tion more closely during early stages than in juve-nile stages, where in the latter the skull and themouth are broad and flat. This does not, however,
212 D. ADRIAENS ET AL.
imply that juvenile and adult C. gariepinus cannotperform any suction. Minute suction forces havebeen demonstrated in ictalurids, which also havedorsoventrally flattened heads, although not to sucha great extent. Broad skull bases also imply thatminute hyoid depressions can generate larger vol-ume expansions (Adriaens and Verraes, 1997e). Thefact that C. gariepinus shows cannibalistic behaviorfrom 8 mm SL on implies that feeding cannot de-pend on suction only, and that grasping is equally oreven more important. Feeding on prey of about thesame size as the predator would require enormoussuction forces if feeding occurred by suction only.
The presence of numerous gill rakers in the juve-nile and adult Clarias gariepinus may be a reflectionof an opportunistic feeding behavior rather than anadaptation to a restricted filter feeding.
Ontogenetic Shift in Functional Demands
It can be predicted that the survival of fish larvaewould be enhanced if the shift in mouth openingmechanisms and the shift in feeding types would becoupled to the shift in functional demands withwhich the developing larva has to deal. As men-tioned above, an ontogenetic shift in diet is presentin most teleosts, as a response to increasing nutri-tional demands. As observed in some cichlids, thetiming of the shift between two mechanisms can beso acute that within 1 day the larva is doomed to dieof starvation or to be able to feed (Otten, 1982). Itwould, however, be advantageous that a safety fac-tor became incorporated in order to reduce suchvulnerable moments during ontogeny (Galis et al.,1994). This can be done in different ways: 1) theelements of a certain apparatus can be formed priorto the moment that the involved mechanism andfunction is needed, or 2) the functioning of differentmechanisms can overlap chronologically: when onemechanism becomes inoperative, the functional de-mand can still be dealt with by the other. For exam-ple, although functional demands for feeding andrespiration are absent in viviparous embiotocids(feeding is intraovarian, whereas respiration occursthrough enlarged and highly vascularized medianfins), a synchronous shift occurs when the protractorhyoidei becomes a mouth closer and the opercularmouth opening mechanism becomes functional(Liem, 1991). In this case, a new reproductive strat-egy reduces the chances of mortality because of pos-sible asynchronies between ontogenetic functionalshifts of cranial elements.
Mouth opening will be important for feeding andrespiratory movements during ontogeny. In thisstudy, we attempted to investigate to what extent arelationship can be found between the timing of thedifferent mouth opening mechanisms during ontog-eny and some morphological, behavioral, or physio-logical changes related to nutritional or respiratorydemands (Fig. 11). Based on these data, five main
phases can be recognized in the life history of Clar-ias gariepinus.
Phase 1 (0–5.0 mm SL). Initially, at hatching nomouth opening is possible. Respiration must occurthrough cutaneous diffusion, as no sign of gills isobserved. A so-called hyoidean vascular net mayfunction as a respiratory organ (Greenwood, 1955).The ventilation of the boundary layer of water, sur-rounding the larva, occurs through undulatorymovements of the notochord, which results in for-ward locomotion. Nutrition is completely dependenton the yolk sac (Hecht and Appelbaum, 1987; Kam-ler et al., 1994).
Phase 2 (5.0–6.2 mm SL). The first movementsof the lower jaw are observed: mouth closing occursthrough contraction of the adductor mandibulae,whereas Mechanism 1 and subsequently Mecha-nism 2 can allow mouth opening. Consequently, thefirst volume changes of the buccal cavity may occur.At this moment the gill filaments have started toform, indicating a possible respiratory shift. Theformation of the opercular cover, as well as its initialmovements, may take part in the initial respiratorypumping mechanism. As the yolk sac is still sub-stantial at this stage, the movements of the mouthpresumably are mainly for respiratory purposes.However, this phase spreads close to the end of theyolk sac period, indicating that active feeding willbecome necessary at the end of this phase. This yolksac phase also allows a learning process, in order toimprove mouth opening and prey capture (Coughlin,1994).
Phase 3 (6.2–11.5 mm SL). This phase starts atthe transition from endogenous to exogenous feed-ing. As the yolk sac period is terminating, adapta-tions to active feeding become essential. The onset ofthis phase is characterized by the formation of thehyoid mouth opening mechanisms (Mechanisms 3and 4) (6.2 mm SL). At this stage, teeth have formed,indicating the potential for prey capture. At 3 daysof age posthatching (6.5–7.0 mm SL), oxygen con-sumption reaches a maximal peak for yolk sac lar-vae of Clarias gariepinus raised at 25°C (Kamler etal., 1994). Respiratory requirements will conse-quently become more important. At 8 mm length, C.gariepinus larvae in captivity show a cannibalisticbehavior, which implies the ability to capture andengulf large prey items (in the natural habitat thiswould mean other small fish species), as well as togenerate large biting forces (cannibalism type I ofHecht and Appelbaum, 1987, 1988). This indicatesthat extensive mouth opening is essential and that apowerful adductor mandibulae is needed. At the endof this phase the opercular mouth opening mecha-nism may become functional, presumably as a prep-aration to fully carnivorous feeding behavior.
Phase 4 (11.5–20 mm SL). This phase is charac-terized by the digestive system, which now becomescompletely functional. At 11.5 mm SL, the stomachhas formed a clear pylorus, functional glandular
213MOUTH OPENING MECHANISMS IN A CATFISH
cells, and pH has dropped below 5 (Stroband andKroon, 1981). As mentioned before, all mouth open-ing mechanisms, which are present in the juveniles,are now present. This means that piscivorous behav-ior can now be fully exploited.
Phase 5 (20–… mm SL). The last phase is char-acterized by the shift in respiratory mechanisms. At20 mm length the suprabranchial organ is formed(Haylor and Oyegunwa, 1993). This structural inno-vation, typical of clariids, enables them to performaerial respiration (Greenwood, 1961). However, theapparatuses present at that stage appear to be suf-ficient to allow the required respiratory movements(Hellin and Chardon, 1983; Vandewalle and Char-don, 1991) as no major changes in related structurescan be observed.
Evolutionary Shift in Mouth OpeningMechanisms
Finally, a brief comment concerning the evolution-ary background of mouth opening mechanisms isoffered. The hyoid mechanism, in which hyoid de-pression is ligamentously coupled to mouth opening,is suggested to be a synapomorphic feature of Te-leostomi (Lauder and Liem, 1980; Lauder, 1980a). Itis only from the halecostomian level on that theopercular mechanism takes part in mouth opening(Lauder, 1980a; Liem, 1991). The increasing impor-tance of the opercular bone is also suggested bymodifications that increase its mobility from thehalecostomian level on (Schaeffer and Rosen, 1961).Apparently, the ontogenetic shift in mouth openingmechanisms is reflected in the evolutionary trendfor improving mouth opening.
CONCLUSIONS
In relation to their accelerated development, earlystages of Clarias gariepinus appear to be a compro-mise between the availability of cranial structuresand functional demands that may have to be copedwith. The rudimentary cranial morphology appearsto be sufficient to form a mouth opening mechanismwhich may sustain the respiratory requirements,whereas the timing of yolk sac depletion appears tobe related to the formation of improved mouth open-ing mechanisms. It is highly probable that the sub-sequent shift and protagonist action of mouth open-ing mechanisms is, to some degree, related to theshift in feeding mode: suction feeding vs. grasping.The increase in functional demands of feeding andrespiratory requirements is reflected in the increasein complexity of the related apparatuses. In C.gariepinus, no critical periods that would requiresynchronous shifts in mouth opening mechanismsappear to be present.
ACKNOWLEDGMENTS
The authors thank F. Ollevier and F. Volckaert(KUL) for the material of the ontogenetic series, andMr. Fleure (the Netherlands) for the juvenile speci-mens. We thank G. De Wever and D. Vandenbroeckfor making the serial sections.
LITERATURE CITEDAdriaens D, Verraes W. 1994. On the functional significance of
the loss of the interhyal during ontogeny in Clarias gariepinusBurchell, 1822 (Teleostei : Siluroidei). Belg J Zool 124:139–155.
Adriaens D, Verraes W. 1996. Ontogeny of cranial musculature inClarias gariepinus (Siluroidei: Clariidae): the adductor man-dibulae complex. J Morphol 229:255–269.
Adriaens D, Verraes W. 1997a. The ontogeny of the chondrocra-nium in Clarias gariepinus: trends in siluroids. J Fish Biol50:1221–1257.
Adriaens D, Verraes W. 1997b. Ontogeny of the hyoid muscula-ture in the African catfish, Clarias gariepinus (Burchell, 1822)(Siluroidei: Clariidae). Zool J Linn Soc 121:105–128.
Adriaens D, Verraes W. 1997c. Ontogeny of the maxillary barbelmuscles in Clarias gariepinus (Siluroidei: Clariidae), with somenotes on the palatine-maxillary mechanism. J Zool Lond 241:117–133.
Adriaens D, Verraes W. 1997d. Ontogeny of the suspensorial andopercular muscles in Clarias gariepinus (Siluroidei: Clariidae),and the consequences for respiratory movements. Neth J Zool47:61–89.
Adriaens D, Verraes W. 1997e. Some consequences of transfor-mations in siluriform chondrocrania: a case study of Clariasgariepinus (Burchell, 1822) (Siluriformes: Clariidae). Neth JZool 47:349–363.
Adriaens D, Verraes W. 1998. Ontogeny of the osteocranium inthe African catfish, Clarias gariepinus (1822) (Siluriformes:Clariidae): ossification sequence as a response to functionaldemands. J Morphol 235:183–237.
Adriaens D, Verraes W, Taverne L. 1997. The cranial lateral-linesystem in Clarias gariepinus (Burchell, 1822) (Siluroidei: Clari-idae): morphology and development of canal related bones. EurJ Morphol 35:181–208.
Aerts P. 1987. Aspecten Van De Voedelopname Bij Beenvissen:Functioneel-Anatomisch Onderzoek Van Astatotilapia Elegans(Trewavas, 1933) (Teleostei: Cichlidae). PhD thesis, Ghent.
Aerts P, Verraes W. 1984. Theoretical analysis of a planar fourbar system in the teleostean skull: the use of mathematics inbiomechanics. Ann Soc R Zool Belg 114:273–290.
Aerts P, Verraes W. 1987. Do inertial effects explain maximalrotation of the maxilla in the rainbow trout (Salmo gairdneri)during feeding? Ann Soc R Zool Belg 117:221–235.
Aerts P, Vanden Berghe W, Verraes W. 1985. Angulo-articularinclination and fast jaw depression in Astatotilapia elegans(Trewavas): different ways of using strain energy. Proc V CongrEur Ichthyol 97–100.
Aerts P, Osse JWM, Verraes W. 1987. Model of jaw depressionduring feeding in Astatotilapia elegans (Teleostei: Cichlidae):mechanisms for energy storage and triggering. J Morphol 194:85–109.
Alexander RM. 1965. Structure and function in the catfish. J Zool(Lond) 148:88–152.
Alexander RM. 1967. The functions and mechanisms of the pro-trusible upper jaws of some acanthopterygian fish. J Zool(Lond) 151:43–64.
Alexander RM. 1970. Mechanics of the feeding action of variousteleost fishes. J Zool (Lond) 162:145–156.
Anker GCh. 1974. Morphology and kinetics of the head of thestickleback, Gasterosteus aculeatus. Trans Zool Soc Lond 32:311–416.
Cook A. 1996. Ontogeny of feeding morphology and kinematics injuvenile fishes: a case study of the cottid fish Clinocottus analis.J Exp Biol 199:1961–1971.
De La Hoz EU, Aldunate R. 1994. El sistema hioideo-mandibularde Cheiorodon (Ostariophysi, Characidae): una innovacion fun-cional. Ann Mus Nat Valparaiso 22:83–90.
de Visser J, Barel CDN. 1996. Architectonic constraints on thehyoid’s optimal starting position for suction feeding of fish. JMorphol 228:1–18.
Drost MR, Muller M, Osse JWM. 1988. A quantitative hydrody-namical model of suction feeding in larval fishes: the role offrictional forces. Proc R Soc Lond (B) 234:263–281.
Elshoud GCA. 1978. Respiration in the three-spined stickleback,Gasterosteus aculeatus L.; an electromyographic approach.Neth J Zool 28:524–544.
Elshoud-Oldenhave MJW, Osse JWM. 1976. Functional morphol-ogy of the feeding system in the ruff — Gymnocephalus cernua(L. 1758) — (Teleostei, Percidae). J Morphol 150:399–422.
Fuiman LA. 1997. What can flatfish ontogenies tell us aboutpelagic and benthic lifestyles? J Sea Res 37:257–267.
Galis F, de Jong PW. 1988. Optimal foraging and ontogeny: foodselection by Haplochromis piceatus. Oecologia 75:175–184.
Galis F, Terlouw A, Osse JWM. 1994. The relation between mor-phology and behaviour during ontogenetic and evolutionarychanges. J Fish Biol 45(Suppl. A):13–26.
Ghiot F, Vandewalle P, Chardon M. 1984. Comparaison anato-mique et fonctionnelle des muscles et des ligaments en rapportavec les barbillons chez deux familles apparentees de poissonsSiluriformes Bagroidei. Ann Soc Roy Zool Belg 114:261–272.
Gill AB. 1997. The role of the mouth morphology in determiningthreespine stickleback (Gasterosteus aculeatus) feeding behav-iour. J Morphol 232:258.
Greenwood PH. 1955. Reproduction in the cat-fish, Clariasmossambicus Peters. Nature 176:516–518.
Greenwood PH. 1961. A revision of the genus Dinotopterus Blgr.(Pisces, Clariidae) with notes on the comparative anatomy ofthe suprabranchial organs in the Clariidae. Bull Br Mus NatHist (Zool) 7:217–241.
Groenewald AAvJ. 1964. Observations on the food habitats ofClarias gariepinus Burchell, the South African freshwater bar-bel (Pisces: Clariidae) in Transvaal. Hydrobiologia 23:287–291.
Haylor GS, Oyegunwa O. 1993. Onset of air breathing and devel-opment of accessory breathing organs in relation to tempera-ture in the African catfish, Clarias gariepinus (Burchell).Aquacult Fish Manag 24:253–260.
Hecht T, Appelbaum S. 1987. Notes on the growth of Israelisharptooth catfish (Clarias gariepinus) during the primarynursing phase. Aquaculture 63:195–204.
Hecht T, Appelbaum S. 1988. Observations on intraspecific ag-gression and coeval sibling cannibalism by larval and juvenileClarias gariepinus (Clariidae: Pisces) under controlled condi-tions. J Zool (Lond) 214:21–44.
Hellin B, Chardon M. 1983. Observations sur le traject de l’airdurant la respiration aerienne chez Clarias lazera Cuvier etValenciennes, 1840. Ann Soc R Zool Belg 113:97–106.
Holden KK, Bruton MN. 1994. The early ontogeny of the mouth-brooder, Pseudocrenilabrus philander (Pisces, Cichlidae). Envi-ron Biol Fish 41:311–329.
Hunt von Herbing I. 1997. Ontogeny of feeding mechanisms inlarval fish with different life histories: winter flounder versusAtlantic cod. J Morphol 232:267.
Hunt von Herbing I, Miyake T, Hall BK, Boutilier RG. 1996a.Ontogeny of feeding and respiration in larval Atlantic cod Ga-dus morhua (Teleostei, Gadiformes). II. Function. J Morphol227:37–50.
Hunt von Herbing I, Miyake T, Hall BK, Boutilier RG. 1996b.Ontogeny of feeding and respiration in larval Atlantic cod Ga-dus morhua (Teleostei, Gadiformes). I. Morphology. J Morphol227:15–35.
Johnson GD. 1992. Monophyly of the euteleostean clades —Neoteleostei, Eurypterygii, and Ctenosquamata. Copeia 1:8–25.
Kamler E, Szlaminska M, Kuczynski M, Hamackova J, Kouril J,Dabrowski R. 1994. Temperature-induced changes of early de-
velopment and yolk utilization in the African catfish Clariasgariepinus. J Fish Biol 44:311–326.
Kohno H, Ordonio-Aguilar R, Ohno A, Taki Y. 1996a. Morpholog-ical aspects of feeding and improvement in feeding ability inearly stage larvae of the milkfish, Chanos chanos. Ichthyol Res43:133–140.
Kohno H, Ordonio-Aguilar R, Ohno A, Taki Y. 1996b. Osteologicaldevelopment of the feeding apparatus in early stage larvae ofthe seabass, Lates calcarifer. Ichthyol Res 43:1–9.
Lauder GV. 1979. Feeding mechanics in primitive teleosts and inthe halecomorph fish Amia calva. J Zool (Lond) 187:543–578.
Lauder GV. 1980a. Evolution of the feeding mechanism in prim-itive actinopterygian fishes: a functional anatomical analysis ofPolypterus, Lepisosteus and Amia. J Morphol 163:283–317.
Lauder GV. 1980b. The role of the hyoid apparatus in the feedingmechanism of the coelacanth Latimeria chalumnae. Copeia1:1–9.
Lauder GV. 1981. Intraspecific functional repertoires in the feed-ing mechanism of the Characoid fishes Lebiasina, Hoplias andChalceus. Copeia 1:154–168.
Lauder GV. 1982. Patterns of evolution in the feeding mechanismof actinopterygian fishes. Am Zool 22:275–285.
Lauder GV, Liem KF. 1980. The feeding mechanism and cephalicmyology of Salvelinus fontenalis: form, function, and evolution-ary significance. In: Balon EK, editor. Charrs, salmonid fishesof the genus Salvelinus. Den Hague, Netherlands: Dr. W. Junk.pp 365–390.
Lauder GV, Liem KF. 1989. The role of historical factors in theevolution of complex organismal functions. In: Wake DB, RothG, editors. Complex organismal functions: integration and evo-lution in vertebrates. Life Sciences Research Report 45. NewYork: John Wiley & Sons. p 63–78.
Lauder GV, Crompton AW, Gans C, Hanken J, Liem KF, MaierWP, Meyer A, Presley R, Rieppel OC, Roth G, Schluter D,Zweers GA. 1989. Group report: how are feeding systems inte-grated and how have evolutionary innovations been intro-duced? In: Wake DB, Roth G, editors. Complex organismalfunctions: integration and evolution in vertebrates. Life Sci-ences Research Report 45. New York: John Wiley & Sons. p97–115.
Liem KF. 1984. The muscular basis of aquatic and aeral ventila-tion in the air-breathing teleost fish Channa. J Exp Biol 113:1–18.
Liem KF. 1991. A functional approach to the development of thehead of teleosts: implications on constructional morphology andconstraints. In: Schmidt-Kittler N, Vogel K, editors. Construc-tional morphology and evolution. Berlin: Springer-Verlag. p231–249.
Lowe CG, Wetherbee BM, Crow GL, Tester AL. 1996. Ontogeneticdietary shifts and feeding behavior of the tiger shark, Galeo-cerdo cuvieri, in Hawaiian waters. Environ Biol Fish 47:203–211.
Muller M. 1987. Optimization principles applied to the mecha-nism of neurocranium levation and mouth bottom depression inbony fishes (Halecostomi). J Theor Biol 126:343–368.
Muller M, Osse JWM. 1984. Hydrodynamics of suction feeding infish. Trans Zool Soc Lond 37:51–135.
Olson MH. 1996. Ontogenetic niche shifts in largemouth bass:variability and consequences for first-year growth. Ecology 77:179–190.
Osse JWM. 1969. Functional morphology of the head of the perch(Perca fluviatilis L.): an electromyographic study. Neth J Zool19:289–392.
Osse JWM. 1989. A functional explanation for a sequence ofdevelopmental events in the carp. The absence of gills in earlylarvae. Acta Morphol Neerland Scand 27:111–118.
Osse JWM. 1990. Form changes in fish larvae in relation tochanging demands of function. Neth J Zool 40:362–385.
Osse JWM, Drost MR. 1989. Hydrodynamics and mechanics offish larvae. Polsk Arch Hydrobiol 36:455–465.
Osse JWM, van den Boogaart JGM. 1995. Fish larvae, develop-ment, allometric growth, and the aquatic environment. ICESMar Sci Symp 201:21–34.
215MOUTH OPENING MECHANISMS IN A CATFISH
Osse JWM, van den Boogaart JGM, van Snik GMJ, van der SluysL. 1997. Priorities during early growth of fish larvae. Aquacul-ture 155:249–258.
Otten E. 1982. The development of a mouth-opening mechanismin a generalized Haplochromis species: H. elegans Trewavas1933 (Pisces, Cichlidae). Neth J Zool 32:31–48.
Radermaker F, Surlemont C, Sanna P, Chardon M, VandewalleP. 1989. Ontogeny of the Weberian apparatus of Clariasgariepinus (Pisces, Siluriformes). Can J Zool 67:2090–2097.
Rowe DK, Chisnall BL. 1996. Ontogenetic habitat shifts by Gal-axias galaxias (Galaxiidae) between the littoral and limneticzones of Lake Kanono, New Zealand. Environ Biol Fish 46:255–264.
Schaefer SA. 1987. Osteology of Hypostomus plecostomus (Lin-naeus) with a phylogenetic analysis of the loricariid subfamilies(Pisces: Siluroidei). Contrib Sci 394:1–31.
Schaefer SA, Lauder GV. 1996. Testing historical hypotheses ofmorphological change: biomechanical decoupling in loricarioidcatfishes. Evolution 50:1661–1675.
Schaeffer B, Rosen DE. 1961. Major adaptive levels in the evolu-tion of the actinopterygian feeding mechanism. Am Zool 1:187–204.
Segnini S, Bastardo H. 1995. Cambios ontogeneticos en la dietade la trucha arcoiris (Oncorhynchus mykiss) en un Rio Andinoneotropical. Biotropica 27:495–508.
Stiassny MLJ. 1996. Basal ctenosquamate relationships and theinterrelationships of the myctophiform (scopelomorph) fishes.In: Stiassny MLJ, Parenti LR, Johnson GD, editors. Interrela-tionships of fishes. London: Academic Press. p 405–426.
Stroband HWJ, Kroon AG. 1981. The development of the stomachin Clarias lazera and the intestinal absorption of protein mac-romolecules. Cell Tissue Res 215:397–415.
Surlemont C, Vandewalle P. 1991. Developpement postembryon-naire du squelette et de la musculature de la tete de Clariasgariepinus (Pisces, Siluriformes) depuis l’eclosion jusqu’a 6,8mm. Can J Zool 69:1094–1103.
Surlemont C, Chardon M, Vandewalle P. 1989. Skeleton, mus-cles and movements of the head of a 5,2 mm fry of Clarias
Teugels GG. 1986. A systematic revision of the African species ofthe genus Clarias (Pisces, Clariidae). Ann Mus R Afr Centr SciZool 247:1–199.
Thomas JD. 1966. On the biology of the catfish Clarias senega-lensis, in a man-made lake in the Ghanaian savanna withparticular reference to its food habits. J Zool (Lond) 148:476–514.
Vandewalle P, Chardon M. 1991. A new hypothesis on the air flowin air breathing in Clarias gariepinus (Teleostei, Siluriformes).Belg J Zool 121:73–80.
Vandewalle P, Surlemont C, Sanna P, Chardon M. 1985. Inter-pretation fonctionnelle de modifications du splanchnocranependant le developpement post-embryonnaire de Clariasgariepinus (Teleosteens, Siluriformes). Zool Jb Anat 113:91–100.
Verraes W. 1977. Postembryonic ontogeny and functional anat-omy of the ligamentum mandibulo-hyoideum and the ligamen-tum interoperculo-mandibulare, with notes on the opercularbones and some other cranial elements in Salmo gairdneriRichardson, 1836 (Teleostei: Salmonidae). J Morphol 151:111–120.
Wainwright PC, Turingan RG. 1993. Coupled versus uncoupledfunctional systems: motor plasticity in the queen triggerfishBalistes vetula. J Exp Biol 180:209–227.
Weibel ER, Taylor CR. 1981. Design of the mammalian respira-tory system. Resp Physiol 44:1–164.
Westneat MW. 1990. Feeding mechanics of teleost fishes (Labri-dae; Perciformes): a test of four-bar linkage models. J Morphol205:269–295.
Westneat MW. 1994. Transmission of force and velocity in thefeeding mechanisms of labrid fishes (Teleostei, Perciformes).Zoomorphology 114:103–118.
Westneat MW, Wainwright PC. 1989. Feeding mechanism ofEpibulus insidiator (Labridae; Teleostei): evolution of a novelfunctional system. J Morphol 202:129–150.
